THE SYSTEMATICS OF CERATIAS 21 



by the difference in length of their dorsal tentacles, and Bertelsen, by his work on the adjustability of 

 the lure, has shown that this is not a real distinction. However, he has separated them from Ceratias 

 holbolli because they comprise the only specimens of Ceratias hitherto caught in southern waters beyond 

 the tropics, and because they have two filaments on the illicial esca. 



There are, at present, three individuals of Bertelsen 's C. tentaculatus. They are the holotype of 

 Mancalias tentaculatus (standard length 7-8 cm.), the type of M. bifilis (7.5 cm.), and the small fish 

 i-6 cm. which Regan & Trewavas (1932, p. 100) have stated ' may belong to this species '. Of these the 

 first and third are preserved in the Collections of the British Museum (Natural History), where I have 

 been able to examine them. In Table 3 and Fig. 5 they are included as nos. 5 and 1 respectively. The 

 type of M. bifilis had been returned to Copenhagen and was not available to me. According to Bertelsen 's 

 classification, the new Antarctic specimen (no. 9) would be regarded as the adult of C. tentaculatus, the 

 other individuals being juveniles. 



In discussing the validity of this species, it may be said, at once, that in the two small individuals 

 (nos. 1 and 5) there are no differences from C. holbolli, excepting perhaps the number of filaments on 

 the esca, which cannot be attributed to juvenility. The fate of their exposed eyes and tubular nostrils, 

 undergoing marked negative allometry, has already been noticed, together with the conformity of their 

 other bodily proportions to the simple allometry equation. Of other features, the pectoral fin in each 

 specimen has seventeen rays. Kroyer's type of C. holbolli and the Antarctic specimen had nineteen 

 rays in the pectoral fin, Regan's specimen eighteen and Bertelsen 's two specimens had seventeen each. 



As in the adult fishes, the number of teeth varies, being — . — in the specimen of Mancalias bifilis and 

 c. — . — in M. tentaculatus, and this is not considered significant. The lateral line organs on the head of 



IS 15 



the Antarctic specimen are found alsojin M. tentaculatus, where they are more extensive, continuing above 

 the eye and then down the body as far as the caudal peduncle. They were not observed in the much 

 smaller specimen of M. bifilis. 



Bertelsen 's first reason for recognizing Ceratias tentaculatus as a new species is that the three 

 individuals comprising its synonyms, Mancalias tentaculatus and M. bifilis, are the only three specimens 

 of Ceratias hitherto taken in waters south of the tropics. So few specimens of Ceratias are known that 

 this implication of a discontinuous distribution is neither supported nor disproved by his chart 

 (Bertelsen, 1943, p. 202), which, equally well, suggests that specimens of the genus may be distributed 

 without interruption through the oceans. 



The morphological feature of an esca bearing two filaments, supposed by Bertelsen to be a second 

 character distinguishing the southern form, is certainly present in the new adult specimen from the 

 Antarctic (Fig. 4C), and in the types of Mancalias tentaculatus (Fig. 4B) and M. bifilis. But in the 

 remaining (fourth) southern specimen, viz. the smaller specimen of M. bifilis (no. 1), there is only 

 a slight rudiment to represent the filament or filaments (Fig. 4A). The esca of this specimen resembles 

 that of the juveniles ' M. sessilis' and ' M. uranoscopus'. Bertelsen, who attributed the naked esca of 

 M. uranoscopus to juvenility, requires that in his classification the northern form should possess only 

 one escal filament in the adult. But of the thirteen known adults of Ceratius holbolli from northern 

 waters, the escas of only two are definitely known to bear one filament, viz. Kroyer's type and Bertelsen 's 

 Greenland specimen (Bertelsen, 1943). In all the other northern adults the cephalic tentacle has either 

 been damaged, lost or not described. 



It is not, therefore, considered that the descriptions and material at present available are sufficient 

 to justify any splitting of the single species apart from the tentative isolation of the anomalous xenistius 

 as a subspecies. It may be, indeed, that Bertelsen is correct and that the escal filaments are a sub- 



